Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A system for identifying healthiness of a plant, comprising: a mobile device having an image sensor for capturing an image of at least a part of a plant and a presentation unit; a database which stores a plurality of preset models; a processing unit adapted to perform an image processing analysis during which a visual representation of said at least part of a plant is extracted from said image, said visual representation is matched against at least one preset model of said plurality of preset models based on a visual similarity, and at least one plant disease is automatically identified based on said at least one preset model; and wherein said presentation unit presents, as an outcome of said image processing analysis, an indication of an absence or a presence of said at least one plant disease; wherein said visual similarity denotes a measure of resemblance between said visual representation and said at least one preset model.
A system for identifying plant health uses a mobile device to capture an image of a plant. The mobile device has an image sensor and a display. A database stores multiple "preset models" of plants, representing different health states or diseases. A processor performs image analysis to extract a visual representation of the plant from the image. This visual representation is compared against the preset models to find the closest visual similarity. Based on the best-matching preset model, the system automatically identifies if a plant disease is present. The mobile device's display then shows an indication of whether the plant disease is present or absent. The "visual similarity" is a measure of how much the captured image resembles a preset model.
2. The system of claim 1 , wherein said outcome is a measure of a fit between the at least one preset model and a specified status of said plant.
The plant health system from the previous description shows an outcome that's a measure of "fit" between a stored preset model and the plant's current condition as captured in the image. So, instead of a simple "disease present/absent", the display could show a numerical score or a percentage indicating how closely the plant matches a healthy or diseased model. This gives a more granular assessment of the plant's health.
3. The system of claim 1 , wherein said at least one preset model comprises a healthy model and a sick model of said plant.
In the plant health system from the first description, the "preset models" stored in the database include at least two types of models: a "healthy model" representing a healthy version of the plant and a "sick model" representing a diseased version of the plant. The system compares the captured image of the plant against both healthy and sick models to determine the closest match and thus identify the health status.
4. The system of claim 1 , wherein said processing unit and said database are integrated into said mobile device.
In the plant health system described earlier, the image processing unit and the database of plant models are both located directly on the mobile device. This means the plant health assessment happens entirely offline, without requiring a network connection. The mobile device captures the image, performs the analysis, and displays the results, all independently.
5. The system of claim 1 , wherein said processing unit and said database are hosted on at least one network node, said mobile device includes a network interface that forwards at least one of said visual representation and said image to said at least one network node and receives said indication from said at least one network node.
In the plant health system described earlier, the image processing unit and the database of plant models are located on a remote server or network node. The mobile device captures the image and sends either the full image or the extracted "visual representation" of the plant to this remote server. The server performs the image analysis, determines the plant's health, and sends back an indication of any diseases. The mobile device then displays this indication. The mobile device communicates with the server through a network interface.
6. The system of claim 1 , wherein said plant is crop.
In the plant health system described earlier, the plant being analyzed is a crop plant. This specifies the system's application domain as agricultural, focusing on identifying diseases or health issues in cultivated crops.
7. The system of claim 1 , wherein said plant is an herb.
In the plant health system described earlier, the plant being analyzed is an herb. This focuses the system on applications in herbal medicine or cultivation of culinary or medicinal herbs.
8. The system of claim 1 , wherein said part is a leaf.
In the plant health system described earlier, the part of the plant being imaged and analyzed is a leaf. The image processing and analysis are specifically tailored to features and characteristics of leaves.
9. The system of claim 8 , wherein said analysis is set to identify a leaf infection.
The plant health system, focused on leaf analysis as mentioned in the previous description, is configured to specifically identify leaf infections. The image processing algorithms and preset models are designed to detect signs of fungal, bacterial, or viral infections on leaves.
10. The system of claim 1 , wherein further comprising means to attach it to a human body.
In the plant health system described earlier, the system includes a way to attach it to a human body. This implies a wearable or hands-free mode of operation, perhaps with a clip or strap to secure the mobile device.
11. The system of claim 1 , wherein said processing unit performs an analysis of said visual representation in light of a location based analysis of localization data acquired by said mobile device.
In the plant health system from the first description, the processing unit considers location data when analyzing the plant's visual representation. The mobile device's location is determined (e.g., via GPS) and this location information is used in the analysis. This allows the system to factor in geographic factors like climate, regional diseases, or soil conditions to improve the accuracy of the plant health assessment.
12. The system of claim 11 , wherein said localization data is selected from a group consisting of: global positioning system data, triangulation data, Global system for mobile communications (GSM) network data, Wi-Fi network data, and a radio frequency identification data.
The location data used by the plant health system, as mentioned in the previous description, can come from several sources: GPS data, triangulation from cell towers, GSM network information, Wi-Fi network data, or RFID data. The system uses one or more of these technologies to determine its location and incorporate that location data into the plant health analysis.
13. The system of claim 1 , wherein said processing unit performs an analysis of said visual representation in light of an analysis of plant health historical data.
In the plant health system described earlier, the processing unit considers historical plant health data when analyzing the image. This means the system has access to records of past plant diseases or health trends in the area or for that specific plant type. This historical data is used to refine the analysis and improve the accuracy of disease detection.
14. The system of claim 1 , wherein said image is part of video data captured using said image sensor.
In the plant health system from the first description, instead of analyzing a single image, the image sensor captures a video of the plant. The image used for analysis is extracted from this video data stream.
15. The system of claim 14 , wherein said processing is performed while further capturing is performed.
In the plant health system that uses video as described in the previous description, the image processing and analysis happen continuously while the video is still being recorded. The system doesn't need to wait for the entire video to be captured before starting the analysis; it processes the video stream in real-time.
16. The system of claim 1 , wherein said at least one preset model is indicative of an aphid on said part of a plant.
In the plant health system described earlier, at least one of the preset models stored in the database specifically represents the presence of aphids on the plant. The system is designed to recognize visual indicators of aphid infestation.
17. The system of claim 15 , wherein said indication is indicative of an aphid category.
Building on the plant health system that identifies aphids as described earlier, the system's output doesn't just indicate the presence of aphids but also provides a category or classification of the aphid infestation. This could include severity, type of aphid, or stage of infestation.
18. A computerized method for identifying healthiness of a plant, comprising: providing an access to a plurality of preset models of a plant; capturing an image of at least a part of a plant using an image sensor of a mobile device; identifying, using a processing unit of said mobile device, at least one plant disease according to an image processing analysis during which a visual representation of said at least part of a plant is extracted from said image, said visual representation is matched against at least one preset model of said plurality of preset models based on a visual similarity, and at least one plant disease is automatically identified based on said at least one preset model; wherein said visual similarity denotes a measure of resemblance between said visual representation and said at least one preset model; and presenting as an outcome of said image processing analysis, an indication of an absence or a presence of said at least one plant disease.
A computer-implemented method for identifying plant health involves accessing a set of stored plant models. An image of a part of the plant is captured using a mobile device's camera. The mobile device's processor then analyzes the image to detect plant diseases. This involves extracting a visual representation of the plant part from the image, comparing this representation to the stored plant models to find the closest visual match, and automatically identifying a plant disease based on the best-matching model. The "visual similarity" is a measure of how closely the captured image resembles a model. The method then presents an indication of whether a plant disease is present or absent.
19. The method according to claim 18 , wherein said image images a shelf space in a marketplace that includes said plant.
The plant health method from the previous description is used to capture images of plants displayed on shelves in a store. This implies the system is used in a retail environment to assess the health or quality of plants being sold.
20. The method according to claim 18 , wherein said at least one plant disease is a presence of an unwanted herb in a plantation which is at least partly depicted in said visual representation.
In the plant health method described earlier, the plant disease being identified is the presence of unwanted weeds in a plantation. The image processing aims to detect and identify weeds growing amongst desired plants, based on the visual representation extracted from the image.
21. The method according to claim 18 , wherein said analysis is performed in less than 5 seconds.
In the plant health method described earlier, the image analysis and disease identification are performed quickly, specifically in less than 5 seconds. This emphasizes the real-time or near real-time nature of the analysis.
22. The method according to claim 18 , wherein said mobile device is attached to a body part of a human.
In the plant health method described earlier, the mobile device used to capture the image is attached to a body part of a human. This allows for hands-free operation during plant inspection.
23. The method according to claim 18 , wherein said identifying is based on keypoint descriptors.
In the plant health method described earlier, the plant disease identification is based on "keypoint descriptors." These are distinctive features or points in the image that are extracted and used to match the image against the stored plant models. This approach is robust to variations in lighting and viewing angle.
24. The method according to claim 18 , wherein said capturing comprises capturing a plurality of images and selecting said image after a plurality of keypoint descriptors are identified therein.
In the plant health method described earlier, the image capture involves taking multiple images, and the system selects the best image for analysis after identifying a sufficient number of keypoint descriptors in each image. The image with the most identifiable features is used for disease detection.
25. The method according to claim 18 , wherein said indication define at least one symptom of a plant illness.
In the plant health method described earlier, the output indicating plant health defines one or more specific symptoms of the plant's illness. The system provides detailed information about the plant's condition, such as leaf discoloration, spots, or wilting.
26. The method of claim 18 , wherein each of said plurality of preset models includes precalculated keypoint descriptors, and said matching of said visual representation is based on matching scene representation keypoint descriptors extracted from said visual representation to said precalculated keypoint descriptors of said at least one preset model.
In the plant health method, each preset model includes pre-calculated keypoint descriptors. The visual representation from the captured plant image is matched by extracting its own keypoint descriptors and comparing them to those pre-calculated for the models. This significantly speeds up the matching process, allowing faster plant disease identification. The system relies on matching these extracted features to identify the closest matching plant model.
Unknown
December 9, 2014
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.